Transferring SPSS Extensions

classic Classic list List threaded Threaded
6 messages Options
JSB
Reply | Threaded
Open this post in threaded view
|

Transferring SPSS Extensions

JSB
I am trying to install over specific 20 SPSS (v.25) extensions and want to copy this process on three computers (Windows 10). Is there an easy way to transfer SPSS extensions between computers? Or perhaps there is syntax that can be copied and pasted to replicate the process? I tried the SPSS extension bundle tool, but it seemed to be quite time consuming and non-intuitive. I'm wondering if there is a simpler way of doing this.
===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
Reply | Threaded
Open this post in threaded view
|

Re: Transferring SPSS Extensions

Jon Peck
Extensions are just a collection of files (no Registry entries).  Typically an extension has an xml file and a code file or files (.R or .py) and a directory by the same name containing a number of other files.  The dialog box files may be in a different location.

For IBM-distributed extensions, the fastest solution might be just to select them all from the Extensions menu and reinstall, but you can just copy the files.  There might be newer versions on the website, too.

What I do is to set two environment variables to specify the location for extensions.
SPSS_CDIALOGS_PATH=C:\dlgs
SPSS_EXTENSIONS_PATH=C:\extcommon

Then just copying everything in or under these locations (via xcopy or as a zip file including the full path) takes care of it.  So if you get one machine set up this way, you can just copy to the others.

However, R-based extensions usually require some R packages that are installed by a different process.  Without looking inside the .spe file, you wouldn't know which are required beyond the R base, so you might want to just let the procedure fail the first time it is run, which will identify the missing packages, and install them as needed.

You can install an R package for use in Statistics like this.
begin program r.
install.packages("package name")
end program.

Using the environment variables also allows you to  have multiple versions of Statistics on the same computer and have the extensions available to all of them without reinstalling in each one.  Extensions are almost never version specific.

If the computers are all on the same network, you can include a network location in the PATH variables above (use ; separators for multiple locations).  Then installing an extension in that location makes it available to all the networked computers, and you can update all of them simultaneously by updating that location.

On Tue, Dec 12, 2017 at 5:31 PM, Brown Vincent <[hidden email]> wrote:
I am trying to install over specific 20 SPSS (v.25) extensions and want to copy this process on three computers (Windows 10). Is there an easy way to transfer SPSS extensions between computers? Or perhaps there is syntax that can be copied and pasted to replicate the process? I tried the SPSS extension bundle tool, but it seemed to be quite time consuming and non-intuitive. I'm wondering if there is a simpler way of doing this.
===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD



--
Jon K Peck
[hidden email]

===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
JSB
Reply | Threaded
Open this post in threaded view
|

Re: Transferring SPSS Extensions

JSB
Thanks Jon. The syntax you posted returns "not recognized as an SPSS Statistics command" error. But I think the general idea is to simply copy and paste the custom dialogs and extensions folders, even if they are in their default locations, correct? (SHOW EXTPATHS)

Now as far as updating extensions, what is the quickest way to update all of them? I suppose through "Update Extensions" in Exension Hub?

On Tue, Dec 12, 2017 at 5:34 PM, Jon Peck <[hidden email]> wrote:
Extensions are just a collection of files (no Registry entries).  Typically an extension has an xml file and a code file or files (.R or .py) and a directory by the same name containing a number of other files.  The dialog box files may be in a different location.

For IBM-distributed extensions, the fastest solution might be just to select them all from the Extensions menu and reinstall, but you can just copy the files.  There might be newer versions on the website, too.

What I do is to set two environment variables to specify the location for extensions.
SPSS_CDIALOGS_PATH=C:\dlgs
SPSS_EXTENSIONS_PATH=C:\extcommon

Then just copying everything in or under these locations (via xcopy or as a zip file including the full path) takes care of it.  So if you get one machine set up this way, you can just copy to the others.

However, R-based extensions usually require some R packages that are installed by a different process.  Without looking inside the .spe file, you wouldn't know which are required beyond the R base, so you might want to just let the procedure fail the first time it is run, which will identify the missing packages, and install them as needed.

You can install an R package for use in Statistics like this.
begin program r.
install.packages("package name")
end program.

Using the environment variables also allows you to  have multiple versions of Statistics on the same computer and have the extensions available to all of them without reinstalling in each one.  Extensions are almost never version specific.

If the computers are all on the same network, you can include a network location in the PATH variables above (use ; separators for multiple locations).  Then installing an extension in that location makes it available to all the networked computers, and you can update all of them simultaneously by updating that location.

On Tue, Dec 12, 2017 at 5:31 PM, Brown Vincent <[hidden email]> wrote:
I am trying to install over specific 20 SPSS (v.25) extensions and want to copy this process on three computers (Windows 10). Is there an easy way to transfer SPSS extensions between computers? Or perhaps there is syntax that can be copied and pasted to replicate the process? I tried the SPSS extension bundle tool, but it seemed to be quite time consuming and non-intuitive. I'm wondering if there is a simpler way of doing this.
===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD



--
Jon K Peck
[hidden email]


===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
Reply | Threaded
Open this post in threaded view
|

Re: Transferring SPSS Extensions

Jon Peck
The "not recognized" error message means that you did not select and run all the lines from the begin program r. (remember the period) through end program, so Statistics tried to interpret the code as regular Statistics commands.

The Extension Hub is the fastest and easiest way to acquire or update all the extensions, but if you set the environment variables as I described, you will have better control over them for the future.

On Wed, Dec 13, 2017 at 12:32 AM, Brown Vincent <[hidden email]> wrote:
Thanks Jon. The syntax you posted returns "not recognized as an SPSS Statistics command" error. But I think the general idea is to simply copy and paste the custom dialogs and extensions folders, even if they are in their default locations, correct? (SHOW EXTPATHS)

Now as far as updating extensions, what is the quickest way to update all of them? I suppose through "Update Extensions" in Exension Hub?

On Tue, Dec 12, 2017 at 5:34 PM, Jon Peck <[hidden email]> wrote:
Extensions are just a collection of files (no Registry entries).  Typically an extension has an xml file and a code file or files (.R or .py) and a directory by the same name containing a number of other files.  The dialog box files may be in a different location.

For IBM-distributed extensions, the fastest solution might be just to select them all from the Extensions menu and reinstall, but you can just copy the files.  There might be newer versions on the website, too.

What I do is to set two environment variables to specify the location for extensions.
SPSS_CDIALOGS_PATH=C:\dlgs
SPSS_EXTENSIONS_PATH=C:\extcommon

Then just copying everything in or under these locations (via xcopy or as a zip file including the full path) takes care of it.  So if you get one machine set up this way, you can just copy to the others.

However, R-based extensions usually require some R packages that are installed by a different process.  Without looking inside the .spe file, you wouldn't know which are required beyond the R base, so you might want to just let the procedure fail the first time it is run, which will identify the missing packages, and install them as needed.

You can install an R package for use in Statistics like this.
begin program r.
install.packages("package name")
end program.

Using the environment variables also allows you to  have multiple versions of Statistics on the same computer and have the extensions available to all of them without reinstalling in each one.  Extensions are almost never version specific.

If the computers are all on the same network, you can include a network location in the PATH variables above (use ; separators for multiple locations).  Then installing an extension in that location makes it available to all the networked computers, and you can update all of them simultaneously by updating that location.

On Tue, Dec 12, 2017 at 5:31 PM, Brown Vincent <[hidden email]> wrote:
I am trying to install over specific 20 SPSS (v.25) extensions and want to copy this process on three computers (Windows 10). Is there an easy way to transfer SPSS extensions between computers? Or perhaps there is syntax that can be copied and pasted to replicate the process? I tried the SPSS extension bundle tool, but it seemed to be quite time consuming and non-intuitive. I'm wondering if there is a simpler way of doing this.
===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD



--
Jon K Peck
[hidden email]


===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD



--
Jon K Peck
[hidden email]

===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
JSB
Reply | Threaded
Open this post in threaded view
|

Re: Transferring SPSS Extensions

JSB
I should have clarified: the error message appeared for the environment variables syntax.

Thank you, your method worked as I was able to transfer the folders between computers (despite being unable to do the environment variable setup). Much easier than having to install them each time.

Do you know if the same easy method can be used with r packages, or is it not recommended to transfer them?

By the way, if there was a way to install extensions via syntax, I think that would be one of the easiest options since we could maintain a list of extensions we want and then just run the same syntax for each SPSS setup. The R packages would be installed automatically as well.



On Wed, Dec 13, 2017 at 6:32 AM, Jon Peck <[hidden email]> wrote:
The "not recognized" error message means that you did not select and run all the lines from the begin program r. (remember the period) through end program, so Statistics tried to interpret the code as regular Statistics commands.

The Extension Hub is the fastest and easiest way to acquire or update all the extensions, but if you set the environment variables as I described, you will have better control over them for the future.

On Wed, Dec 13, 2017 at 12:32 AM, Brown Vincent <[hidden email]> wrote:
Thanks Jon. The syntax you posted returns "not recognized as an SPSS Statistics command" error. But I think the general idea is to simply copy and paste the custom dialogs and extensions folders, even if they are in their default locations, correct? (SHOW EXTPATHS)

Now as far as updating extensions, what is the quickest way to update all of them? I suppose through "Update Extensions" in Exension Hub?

On Tue, Dec 12, 2017 at 5:34 PM, Jon Peck <[hidden email]> wrote:
Extensions are just a collection of files (no Registry entries).  Typically an extension has an xml file and a code file or files (.R or .py) and a directory by the same name containing a number of other files.  The dialog box files may be in a different location.

For IBM-distributed extensions, the fastest solution might be just to select them all from the Extensions menu and reinstall, but you can just copy the files.  There might be newer versions on the website, too.

What I do is to set two environment variables to specify the location for extensions.
SPSS_CDIALOGS_PATH=C:\dlgs
SPSS_EXTENSIONS_PATH=C:\extcommon

Then just copying everything in or under these locations (via xcopy or as a zip file including the full path) takes care of it.  So if you get one machine set up this way, you can just copy to the others.

However, R-based extensions usually require some R packages that are installed by a different process.  Without looking inside the .spe file, you wouldn't know which are required beyond the R base, so you might want to just let the procedure fail the first time it is run, which will identify the missing packages, and install them as needed.

You can install an R package for use in Statistics like this.
begin program r.
install.packages("package name")
end program.

Using the environment variables also allows you to  have multiple versions of Statistics on the same computer and have the extensions available to all of them without reinstalling in each one.  Extensions are almost never version specific.

If the computers are all on the same network, you can include a network location in the PATH variables above (use ; separators for multiple locations).  Then installing an extension in that location makes it available to all the networked computers, and you can update all of them simultaneously by updating that location.

On Tue, Dec 12, 2017 at 5:31 PM, Brown Vincent <[hidden email]> wrote:
I am trying to install over specific 20 SPSS (v.25) extensions and want to copy this process on three computers (Windows 10). Is there an easy way to transfer SPSS extensions between computers? Or perhaps there is syntax that can be copied and pasted to replicate the process? I tried the SPSS extension bundle tool, but it seemed to be quite time consuming and non-intuitive. I'm wondering if there is a simpler way of doing this.
===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD



--
Jon K Peck
[hidden email]


===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD



--
Jon K Peck
[hidden email]


===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
Reply | Threaded
Open this post in threaded view
|

Re: Transferring SPSS Extensions

Jon Peck
Environment variables belong to the os.  They have to be set via the Windows Control Panel.  Then Statistics will use them after it is restarted.

For R packages, you can create a .R file with a list of install.packages calls and run it via
R CMD BATCH mylist.R.

I agree that a batch extension command installer/updated would be a good idea, but currently it does not exist AFAIK.



On Fri, Dec 15, 2017 at 7:33 PM Brown Vincent <[hidden email]> wrote:
I should have clarified: the error message appeared for the environment variables syntax.

Thank you, your method worked as I was able to transfer the folders between computers (despite being unable to do the environment variable setup). Much easier than having to install them each time.

Do you know if the same easy method can be used with r packages, or is it not recommended to transfer them?

By the way, if there was a way to install extensions via syntax, I think that would be one of the easiest options since we could maintain a list of extensions we want and then just run the same syntax for each SPSS setup. The R packages would be installed automatically as well.



On Wed, Dec 13, 2017 at 6:32 AM, Jon Peck <[hidden email]> wrote:
The "not recognized" error message means that you did not select and run all the lines from the begin program r. (remember the period) through end program, so Statistics tried to interpret the code as regular Statistics commands.

The Extension Hub is the fastest and easiest way to acquire or update all the extensions, but if you set the environment variables as I described, you will have better control over them for the future.

On Wed, Dec 13, 2017 at 12:32 AM, Brown Vincent <[hidden email]> wrote:
Thanks Jon. The syntax you posted returns "not recognized as an SPSS Statistics command" error. But I think the general idea is to simply copy and paste the custom dialogs and extensions folders, even if they are in their default locations, correct? (SHOW EXTPATHS)

Now as far as updating extensions, what is the quickest way to update all of them? I suppose through "Update Extensions" in Exension Hub?

On Tue, Dec 12, 2017 at 5:34 PM, Jon Peck <[hidden email]> wrote:
Extensions are just a collection of files (no Registry entries).  Typically an extension has an xml file and a code file or files (.R or .py) and a directory by the same name containing a number of other files.  The dialog box files may be in a different location.

For IBM-distributed extensions, the fastest solution might be just to select them all from the Extensions menu and reinstall, but you can just copy the files.  There might be newer versions on the website, too.

What I do is to set two environment variables to specify the location for extensions.
SPSS_CDIALOGS_PATH=C:\dlgs
SPSS_EXTENSIONS_PATH=C:\extcommon

Then just copying everything in or under these locations (via xcopy or as a zip file including the full path) takes care of it.  So if you get one machine set up this way, you can just copy to the others.

However, R-based extensions usually require some R packages that are installed by a different process.  Without looking inside the .spe file, you wouldn't know which are required beyond the R base, so you might want to just let the procedure fail the first time it is run, which will identify the missing packages, and install them as needed.

You can install an R package for use in Statistics like this.
begin program r.
install.packages("package name")
end program.

Using the environment variables also allows you to  have multiple versions of Statistics on the same computer and have the extensions available to all of them without reinstalling in each one.  Extensions are almost never version specific.

If the computers are all on the same network, you can include a network location in the PATH variables above (use ; separators for multiple locations).  Then installing an extension in that location makes it available to all the networked computers, and you can update all of them simultaneously by updating that location.

On Tue, Dec 12, 2017 at 5:31 PM, Brown Vincent <[hidden email]> wrote:
I am trying to install over specific 20 SPSS (v.25) extensions and want to copy this process on three computers (Windows 10). Is there an easy way to transfer SPSS extensions between computers? Or perhaps there is syntax that can be copied and pasted to replicate the process? I tried the SPSS extension bundle tool, but it seemed to be quite time consuming and non-intuitive. I'm wondering if there is a simpler way of doing this.
===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD



--
Jon K Peck
[hidden email]


===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD



--
Jon K Peck
[hidden email]


===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD
--
Jon K Peck
[hidden email]

===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD